Diabetic Retinopathy Detection Using Retinal Fundus Picture and Image Enhancement Using Fuzzy Clustering

Patients with diabetes are more susceptible to developing diabetic retinopathy. Blindness can be caused by a disease and can be saved if early diabetic retinopathy detection is made. Retinal exudates and haemorrhages associated with diabetic retinopathy can be easily detected in fundus images due to their sensitivity to vascular disorders. Exudates and haemorrhages in retinal fundus images can now be easily and affordably detected, and diabetic retinopathy can be appropriately classified. The fundus image is clustered into different areas of interest using a k-means colour compression technique, which reduces the colour dimension. The region properties attributes were used to separate and categories the various components of the diabetic fundus. Finally, experiment and trial were used to identify diabetic retinopathy using the knowledge-based fuzzy inference system (FIS). FIS recognize diabetic retinopathy by utilizing the above-mentioned effective attributes and knowledge-based fuzzy inference system (FIS). A deep learning-based Efficient Net-based feature extractor is used to produce feature vectors from pre-processed photos. The MESSIDOR dataset is used for all trials, and the outcomes are evaluated using a range of criteria. The K-Means Clustering with Fuzzy Technique approach outperformed the competition in rapports of diagnostic accurateness.